How does one compare the goalies of the Premier League? Based on Opta data released by Manchester City last year, I have compared the goalkeepers of the 2011-12 season on two parameters – percentage of shots blocked and success in distribution. This shows the relative successes of the goalkeepers in defence and attack respectively.
So who are the successful goalkeepers in the league? If you look for at this data, you will find that Pepe Reina, Wojciech Szczesny, Petr Cech and Joe Hart form a “convex hull”. What this means is that every other goalkeeper in the premier league is inferior to at least one of these four. So the best goalie has to be one of these.
As for who the absolute best is, that will depend on the relative weights that we give to shot-stopping and distributional ability. Intuitively, a 10% improvement in distribution is likely to result in less goals than a 10% improvement in shot stopping saves. So by that metric, it can be argued that one of Cech and Hart, who are far superior in terms of shot stopping ability is the best goalkeeper.
It is also interesting to note that even in Kenny Dalglish’s time as manager, Reina had a vastly superior distribution success, suggesting that it would not have been that difficult for Liverpool to adapt to Brendan Rodgers’s style of play.
PS: Click on the image to see a larger version
PS2: I’m writing this post sitting in the office room of one of my clients, who also happens to be a frequent visitor to this blog.
These are nice metrics, considering the fact that the quality of the back four (or three as the case may be) does not affect both these metrics.
I am not really surprised by the top 4 (Szczesny was very good last season) or the next in line – Vorm, Al Habsi, Friedel and Schwarzer.
But Ruddy surprises me. Totally underrated going by these stats. Potential Moneyball signing?
Ruddy was always known to be a good goalie. Was in the England squad for the Euros (second choice to Hart) but then injured his finger and had to pull out. He has also missed most of the season so far.
He might be another Graeme Smith (poppadum fingers) for all you know which will devalue him
Good stuff da. You should do more such stuff. And ofcourse, Hart is the best.:P
How is distribution ability defined? Percentage of passes that reached someone from the same team?
Some further points:
1) The relegated keepers are lower on both scales. (Hennessey, Robinson and Bogdan + Jaskelainen)
2) I expected Vorm to be higher on the distribution scale. He was specifically recruited for this reason.
yeah. percentage of passes that reached same team. And as someone pointed out on twitter, Reina’s high score might be because he usually passes to defenders and never punts. I would also claim that he is a really good really deep lying playmaker and directs attacks!
Vorm is pretty high on the distribution scale. Only Reina and Szczesny are above him. And it would be interesting to analyze how the relegated teams would have done with better goalies (will be my next post)
Can’t understand how/why is distribution ability a measure of the quality of the goal keeper.
because the attack begins from him. Reina for sure has done a great job as a really deep lying playmaker who starts off attacks.
I believe the four goalies form a “Pareto Frontier” which is a subset of the convex hull.
Yup, my apologies! My tech fundaes have become a bit rusty! I knew clearly they are not a dominating set. And yes, they form a part of the convex hull, though